Three powerful tools to transform your mahjong skills, powered by deep reinforcement learning.
Input your current hand and get instant AI recommendations:
Upload your game log and get a comprehensive analysis:
5 hand scenarios daily to sharpen your decision-making:
Our AI is built on state-of-the-art deep reinforcement learning.
Four AI agents play millions of games against each other, exploring strategies no human has tried.
Proximal Policy Optimization refines the AI's strategy, balancing exploration and exploitation.
Potential-based reward shaping guides learning with intermediate signals — not just win/lose.
Models train 24/7 on cloud GPUs, getting stronger every day. Current training: 500K+ episodes per variant.
Free during beta. No download, no signup required to try.